Commit Graph

121294 Commits

Author SHA1 Message Date
Peter Collingbourne
4fda724f03 lit: Match UNSUPPORTED against target triple as we do for XFAIL.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246343 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 22:17:28 +00:00
Vedant Kumar
21f084aa72 [X86] NFC: Clean up and clang-format a few lines
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246340 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 21:59:00 +00:00
Duncan P. N. Exon Smith
819d0a503e DI: Add Function::getSubprogram()
Add `Function::setSubprogram()` and `Function::getSubprogram()`,
convenience methods to forward to `setMetadata()` and `getMetadata()`,
respectively, and deal in `DISubprogram` instead of `MDNode`.

Also add a verifier check to enforce that `!dbg` attachments are always
subprograms.

Originally (when I had the llvm-dev discussion back in April) I thought
I'd store a pointer directly on `llvm::Function` for these attachments
-- we frequently have debug info, and that's much cheaper than using map
in the context if there are no other function-level attachments -- but
for now I'm just using the generic infrastructure.  Let's add the extra
complexity only if this shows up in a profile.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246339 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 21:55:35 +00:00
Duncan P. N. Exon Smith
155b4af95d AsmPrinter: Allow null subroutine type
Currently the DWARF backend requires that subprograms have a type, and
the type is ignored if it has an empty type array.  The long term
direction here -- see PR23079 -- is instead to skip the type entirely if
there's no valid type.

It turns out we have cases in tree of missing types on subprograms, but
since they're not referenced by compile units, the backend never crashes
on them.  One option would be to add a Verifier check that subprograms
have types, and fix the bitrot.  However, this is a fair bit of churn
(20-30 testcases) that would be reversed anyway by PR23079.

I found this inconsistency because of a WIP patch and upgrade script for
PR23367 that started crashing on test/DebugInfo/2010-10-01-crash.ll.
This commit updates the testcase to reference the subprogram from the
compile unit, and fixes the resulting crash (in line with the direction
of PR23079).  This also updates `DIBuilder` to stop assuming a non-null
pointer for the subroutine types.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246333 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 21:38:24 +00:00
David Majnemer
c13175f52f Revert r246232 and r246304.
This reverts isSafeToSpeculativelyExecute's use of ReadNone until we
split ReadNone into two pieces: one attribute which reasons about how
the function reasons about memory and another attribute which determines
how it may be speculated, CSE'd, trap, etc.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246331 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 21:13:39 +00:00
Matthew Simpson
df2dbd7b15 Remove white space (test commit)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246329 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 20:38:33 +00:00
Rafael Espindola
7ef469f951 Split the gold tests into X86 and PowerPC directories.
Patch by Than McIntosh!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246328 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 20:33:56 +00:00
Duncan P. N. Exon Smith
a5ae7c1c9f DI: Require subprogram definitions to be distinct
As a follow-up to r246098, require `DISubprogram` definitions
(`isDefinition: true`) to be 'distinct'.  Specifically, add an assembler
check, a verifier check, and bitcode upgrading logic to combat testcase
bitrot after the `DIBuilder` change.

While working on the testcases, I realized that
test/Linker/subprogram-linkonce-weak-odr.ll isn't relevant anymore.  Its
purpose was to check for a corner case in PR22792 where two subprogram
definitions match exactly and share the same metadata node.  The new
verifier check, requiring that subprogram definitions are 'distinct',
precludes that possibility.

I updated almost all the IR with the following script:

    git grep -l -E -e '= !DISubprogram\(.* isDefinition: true' |
    grep -v test/Bitcode |
    xargs sed -i '' -e 's/= \(!DISubprogram(.*, isDefinition: true\)/= distinct \1/'

Likely some variant of would work for out-of-tree testcases.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246327 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 20:26:49 +00:00
Sanjoy Das
18188f1261 [IR] Add some asserts to CallInst and InvokeInst; NFCI.
The asserts check that accessors supposed to access call / invoke
arguments only ever access call / invoke arguments.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246316 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 19:09:34 +00:00
Sanjoy Das
c93f771c2f [InstCombine] Fix PR24605.
PR24605 is caused due to an incorrect insert point in instcombine's IR
builder.  When simplifying

  %t = add X Y
  ...
  %m = icmp ... %t

the replacement for %t should be placed before %t, not before %m, as
there could be a use of %t between %t and %m.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246315 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 19:09:31 +00:00
Chad Rosier
55189b7fca Optimize memcmp(x,y,n)==0 for small n and suitably aligned x/y.
http://reviews.llvm.org/D6952
PR20673

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246313 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 18:30:18 +00:00
Vedant Kumar
51b3e61e69 [test] (NFC) Simplify Transforms/ConstProp/calls.ll
Differential Revision: http://reviews.llvm.org/D12421

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246312 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 18:04:20 +00:00
Petar Jovanovic
109b926df1 [mips64][mcjit] Add N64R6 relocations tests and fix N64R2 tests
This patch adds a test for MIPS64R6 relocations, it corrects check
expressions for R_MIPS_26 and R_MIPS_PC16 relocations in MIPS64R2 test, and
it adds run for big endian in MIPS64R2 test.

Patch by Vladimir Radosavljevic.

Differential Revision: http://reviews.llvm.org/D11217


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246311 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 18:02:53 +00:00
Petar Jovanovic
ca95e2ee93 [mips] Remove incorrect DebugLoc entries from prologue
This has been causing the prologue_end to be incorrectly positioned.

Patch by Vladimir Radosavljevic.

Differential Revision: http://reviews.llvm.org/D11293


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246309 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 17:53:26 +00:00
Matt Arsenault
1350385078 Make MergeConsecutiveStores look at other stores on same chain
When combiner AA is enabled, look at stores on the same chain.
Non-aliasing stores are moved to the same chain so the existing
code fails because it expects to find an adajcent store on a consecutive
chain.

Because of how DAGCombiner tries these store combines,
MergeConsecutiveStores doesn't see the correct set of stores on the chain
when it visits the other stores. Each store individually has its chain
fixed before trying to merge consecutive stores, and then tries to merge
stores from that point before the other stores have been processed to
have their chains fixed. To fix this, attempt to use FindBetterChain
on any possibly neighboring stores in visitSTORE.

Suppose you have 4 32-bit stores that should be merged into 1 vector
store. One store would be visited first, fixing the chain. What happens is
because not all of the store chains have yet been fixed, 2 of the stores
are merged. The other 2 stores later have their chains fixed,
but because the other stores were already merged, they have different
memory types and merging the two different sized stores is not
supported and would be more difficult to handle.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246307 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 17:31:28 +00:00
David Majnemer
dc15d2d8ea Test case for r246304.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246306 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 17:19:54 +00:00
JF Bastien
1616e55c95 Remove Merge Functions pointer comparisons
Summary:
This patch removes two remaining places where pointer value comparisons
are used to order functions: comparing range annotation metadata, and comparing
block address constants. (These are both rare cases, and so no actual
non-determinism was observed from either case).

The fix for range metadata is simple: the annotation always consists of a pair
of integers, so we just order by those integers.

The fix for block addresses is more subtle. Two constants are the same if they
are the same basic block in the same function, or if they refer to corresponding
basic blocks in each respective function. Note that in the first case, merging
is trivially correct. In the second, the correctness of merging relies on the
fact that the the values of block addresses cannot be compared. This change is
actually an enhancement, as these functions could not previously be merged (see
merge-block-address.ll).

There is still a problem with cross function block addresses, in that constants
pointing to a basic block in a merged function is not updated.

This also more robustly compares floating point constants by all fields of their
semantics, and fixes a dyn_cast/cast mixup.

Author: jrkoenig
Reviewers: dschuff, nlewycky, jfb
Subscribers llvm-commits
Differential revision: http://reviews.llvm.org/D12376

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246305 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 16:49:09 +00:00
David Majnemer
c03acbca56 [CodeGen] isInTailCallPosition didn't consider readnone tailcalls
A readnone tailcall may still have a chain of computation which follows
it that would invalidate a tailcall lowering.  Don't skip the analysis
in such cases.

This fixes PR24613.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246304 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 16:44:09 +00:00
James Molloy
f5ab7a4dbb [LoopUtils] Move a private constructor nearer the other private members
This was part of Adam Nemet's review feedback that I forgot to implement.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246301 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 14:40:29 +00:00
Sanjay Patel
4b1821fa36 [x86] enable machine combiner reassociations for scalar 'and' insts
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246300 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 14:09:48 +00:00
Davide Italiano
a82bdabbf2 [MC] Convert tests to use llvm-readobj --macho-version-min.
As an added bonus this also tests the newly introduced feature.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246296 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 12:40:05 +00:00
Rui Ueyama
c2aa1793ed llvm-readobj: Dump more info for COFF import libraries.
This patch teaches llvm-readobj to print out COFF import file header fields.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246291 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 10:27:50 +00:00
Chandler Carruth
9befb59470 [SROA] Fix PR24463, a crash I introduced in SROA by allowing it to
handle more allocas with loads past the end of the alloca.

I suspect there are some related crashers with slightly different
patterns, but I'll fix those and add test cases as I find them.

Thanks to David Majnemer for the excellent test case reduction here.
Made this super simple to debug and fix.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246289 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 09:03:52 +00:00
Rui Ueyama
3bc806de10 Attempt to unbreak Win32 build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246284 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 07:48:41 +00:00
Rui Ueyama
40e22518fc Re-apply r246276 - Object: Teach llvm-ar to create symbol table for COFF short import files
This patch includes a fix for a llvm-readobj test. With this patch, 
the tool does no longer print out COFF headers for the short import
file, but that's probably desirable because the header for the short
import file is dummy.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246283 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 07:40:30 +00:00
Steven Wu
109cab4450 Revert r246244 and r246243
These two commits cause clang/llvm bootstrap to hang.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246279 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 06:52:00 +00:00
Rui Ueyama
976f3e14bf Rollback r246276 - Object: Teach llvm-ar to create symbol table for COFF short import files
This change caused a test for llvm-readobj to fail.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246277 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 06:03:01 +00:00
Rui Ueyama
e8402a23ca Object: Teach llvm-ar to create symbol table for COFF short import files.
COFF short import files are special kind of files that contains only
DLL-exported symbol names. That's different from object files because
it has no data except symbol names.

This change implements a SymbolicFile interface for the short import
files so that symbol names can be accessed through that interface.
llvm-ar is now able to read the file and create symbol table entries
for short import files.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246276 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 05:47:46 +00:00
NAKAMURA Takumi
d2d41f20f7 LLVMCodeGen: Update libdeps corresponding to r246236.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246274 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 05:38:49 +00:00
Peter Collingbourne
06cc30573e Tweak XFAIL line for mips.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246269 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 04:07:52 +00:00
NAKAMURA Takumi
51c7ae83c6 Kaleidoscope: Prune __attribute__((used)). Some compilers wouldn't accept one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246268 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 03:34:33 +00:00
NAKAMURA Takumi
e8a7097338 Disable llvm/test/Examples/ for now while investigating.
FIXME:
  - Introduce explicit mapping.
  - Investigate crash on win32. Could we introduce crash handler in examples?

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246267 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 03:32:43 +00:00
Peter Collingbourne
ef9fc8314e XFAIL parallel.ll test on MIPS and AArch64 until test failures can be investigated.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246261 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 02:14:15 +00:00
Ahmed Bougacha
5edf06bb57 [CodeGen] Support (and default to) expanding READCYCLECOUNTER to 0.
For targets that didn't support this, this will let us respect the
langref instead of failing to select.

Note that we don't need to change the 32-bit x86/PPC lowerings (to
account for the result type/# difference) because they're both
custom and bypass type legalization.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246258 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 01:49:59 +00:00
Joseph Tremoulet
0bf9bbac45 [WinEH] Update coloring to handle nested cases cleanly
Summary:
Change the coloring algorithm in WinEHPrepare to visit a funclet's exits
in its parents' contexts and so properly classify the continuations of
nested funclets.

Also change the placement of cloned blocks to be deterministic and to
maintain the relative order of each funclet's blocks.

Add a lit test showing various patterns that require cloning, the last
several of which don't have CHECKs yet because they require cloning
entire funclets which is NYI.


Reviewers: rnk, andrew.w.kaylor, majnemer

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D12353

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246245 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 01:12:35 +00:00
Piotr Padlewski
bcd94e99ee Constant propagation after hitting assume(cmp) bugfix
Last time code run into assertion `BBE.isSingleEdge()` in
lib/IR/Dominators.cpp:200.

http://reviews.llvm.org/D12170

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246244 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 01:02:00 +00:00
Piotr Padlewski
0f4937c492 Constant propagation after hiting llvm.assume
After hitting @llvm.assume(X) we can:
- propagate equality that X == true
- if X is icmp/fcmp (with eq operation), and one of operand
  is constant we can change all variables with constants in the same BasicBlock

http://reviews.llvm.org/D11918

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246243 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 01:01:57 +00:00
NAKAMURA Takumi
87a13599bb [CMake] Fix build on MSVC in r246156.
add_windows_version_resource_file() affects ALL_FILES. OBJLIB shouldn't have *.obj as SOURCES.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246241 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 00:36:58 +00:00
George Burgess IV
dc5cb4d4ca Fix: CFLAA -- Mark no-args returns as unknown
Prior to this patch, we hadn't been marking StratifiedSets with the
appropriate StratifiedAttrs when handling the result of no-args call
instructions. This caused us to report NoAlias when handed, for
example, an escaped alloca and a result from an opaque function. Now we
properly mark the return value of said functions.

Thanks again to Chandler, Richard, and Nick for pinging me about this.

Differential review: http://reviews.llvm.org/D12408



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246240 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-28 00:16:18 +00:00
Quentin Colombet
4a93cd8f66 [AArch64][CollectLOH] Fix a regression that prevented us to detect chains of
more than 2 instructions.

I introduced this regression a while back and did not noticed it because I
somehow forgot to push the initial test cases for the pass!

Fix that as well!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246239 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 23:47:10 +00:00
Peter Collingbourne
daf6885fa4 CodeGen: Introduce splitCodeGen and teach LTOCodeGenerator to use it.
llvm::splitCodeGen is a function that implements the core of parallel LTO
code generation. It uses llvm::SplitModule to split the module into linkable
partitions and spawning one code generation thread per partition. The function
produces multiple object files which can be linked in the usual way.

This has been threaded through to LTOCodeGenerator (and llvm-lto for testing
purposes). Separate patches will add parallel LTO support to the gold plugin
and lld.

Differential Revision: http://reviews.llvm.org/D12260

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246236 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 23:37:36 +00:00
Reid Kleckner
c0e64ada5c [WinEH] Add some support for code generating catchpad
We can now run 32-bit programs with empty catch bodies.  The next step
is to change PEI so that we get funclet prologues and epilogues.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246235 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 23:27:47 +00:00
Michael J. Spencer
82f1bf2f60 [yaml2obj] Support numeric indexes to create invalid files. Will be used by lld test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246233 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 23:11:20 +00:00
David Majnemer
2cdf5c3e03 [ValueTracking] readnone CallInsts are fair game for speculation
Any call which is side effect free is trivially OK to speculate.  We
already had similar logic in EarlyCSE and GVN but we were missing it
from isSafeToSpeculativelyExecute.

This fixes PR24601.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246232 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 23:03:01 +00:00
Lang Hames
ef38201ab0 Add a global mapping layer for Orc. Adapted from a patch by Andy Somogyi.
Thanks Andy!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246226 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 22:20:05 +00:00
Peter Collingbourne
a74d34a8a9 Support: Introduce thread.h.
This header is a wrapper for <thread> that works around problems with the
MSVC headers when exceptions are disabled.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246218 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 21:52:31 +00:00
Ahmed Bougacha
69019c5399 [CodeGen] Check FoldConstantArithmetic result before using it.
Fixes PR24602: r245689 introduced an unguarded use of
SelectionDAG::FoldConstantArithmetic, which returns 0 when it fails
because of opaque (hoisted) constants.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246217 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 21:46:04 +00:00
Ahmed Bougacha
bfe0f6957b [TableGen][DAGISel] Dedup predicates with same code to run. NFCI.
I locally hit the 255 limit, but a lot of these are redundant: each
predicate coming from a different record was allocated a new number,
even when we already emitted the same code for another predicate.

Instead, re-use numbers and emit the predicate code only once.
This reduces the total text size of *DAGISel.cpp.o by ~1%.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246208 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 20:43:34 +00:00
Tyler Nowicki
aa08d66b7b Fix test introduced in r246187 that failed on some systems.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246207 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 20:43:29 +00:00
Lang Hames
b6347771b1 Oops - Re-add the Kaleidoscope regression tests themselves (accidentally left
out of r246201).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@246203 91177308-0d34-0410-b5e6-96231b3b80d8
2015-08-27 20:33:22 +00:00